Right hemisphere brain damage (RHD) in adults frequently causes comprehension difficulties. One particular aspect of comprehension, inferencing, has been implicated as a source of these difficulties, although it is not clear what stage(s) of the inferencing process could be impaired. Studies of normal inferencing processes suggest that characteristics of contexts, such as recency of mention of inferences and strength of contextual bias, may affect such inferencing processes. Individual differences in comprehension skill and working memory capacity also play a role. These factors have received little attention in the RHD literature. The long-term objective of the proposed project is to build a theory of inferencing deficits in adults with RHD. These objectives are related to the health and well being of stroke patients, as the results of the study will provide a clearer understanding of RHD communication disorders and may improve the efficacy and efficiency of therapy approaches. The specific aims are to: (a) test competing hypotheses regarding recency of mention effects, such that predictive inferences 'disappear' when tested at a point distant from where they were suggested, (b) examine the influence of contextual bias on predictive inferencing, (c) test the assumption that specific inferencing processes are related to discourse comprehension skill and working memory, and (d) explore brain behavior relationships associated with inferencing processes. Two studies will be conducted to address these aims. Methods: Both studies will employ a line-by-line reading time task using experimental stories constructed to bias toward a specific predictive inference, with a target sentence that disconfirms the intended inference. Slowed reading time on the target sentence in experimental versus control conditions will indicate activation of the intended inference. Study A will address recency of mention effects by manipulating the distance between when an inference is suggested and the point of measurement. Reading times for target and post-target sentences will provide evidence for the source of recency of mention effects. In Study B, strength of contextual bias will be manipulated to examine inferencing under contexts will contain either weak or strong bias toward a specific predictive inference. Data from the reading time studies will be correlated with performance on a general discourse comprehension test and a working memory task to evaluate the relationship between inferencing and these participant characteristics. For both studies, the research design will be a mixed design, with group (RHD and non-brain-damaged) as the between-subjects variable and story condition as the within-subjects variable.